[PATCH v2 3/6] arm: KVM: Invalidate BTB on guest exit

[Marc Zyngier]

On 23/01/18 14:22, Christoffer Dall wrote:
> On Mon, Jan 08, 2018 at 06:55:30PM +0000, Marc Zyngier wrote:
>> In order to avoid aliasing attacks against the branch predictor,
>> let's invalidate the BTB on guest exit. This is made complicated
>> by the fact that we cannot take a branch before invalidating the
>> BTB.
> 
> ...because that would defeat the overall purpose of what we're trying to
> do?

Indeed. Even direct branches could potentially be unsafe, so branching
at that stage is not a good idea.

> 
>>
>> Another thing is that we perform the invalidation on all
>> implementations, no matter if they are affected or not.
> 
> I don't understand this comment, it seems like the logic below is
> limited to A12 and A17 for now?

Ah, that's a leftover from the first version (before I discovered that
A15 needed a different method). I'll amend that when I repost the patches.

Thanks,

	M.
-- 
Jazz is not dead. It just smells funny...